Deployment

VMs are deployed using some form of a hypervisor. With Windows systems, they can be deployed and managed by Windows Admin Center, Hyper-V Manager, and System Center Virtual Machine Manager (SCVMM). PowerShell can also be used with the relevant Hyper-V modules enabled.

Containers, in comparison to VMs, are deployed using Docker through the command line for single instances, and for multiple containers, it is advised that an orchestrator service such as AKS is used.

Takeaway

Understand the tools for each deployment type, knowing that on Azure, we predominantly use Docker containers and AKS for container orchestration.

Storage persistence

VM storage is very easy to scale and maintain in Azure. We simply add or scale up disks as required, as you experienced in the previous chapter. This storage is persistent as it doesn’t disappear if disconnected or if the VM shuts down. Another mechanism of providing persistent storage is through Server Message Block (SMB) shares.

Containers, in comparison, have the option of leveraging Azure disks for local storage on a single-node deployment and configuring SMB shares for multiple nodes or servers.

Takeaway

Identify whether storage requires persistence and find the mechanism that best suits your requirements. For shared storage, SMB shares make a lot of sense, but there are several other mechanisms that can be used to achieve the same result. These are beyond the scope of this book, however, but you are still encouraged to explore and find what works for you.

Fault tolerance

VMs can fail over to another server in a cluster when configured for fault tolerance. The resultant effect, however, is that the VM’s OS will be restarted on the new hosting server, which can add potentially significant time to restore services.

Containers, in comparison, have the ability to quickly recover instances. The orchestrator recreates the container on another cluster node when it detects a failed cluster node.

Takeaway

Catering to fault tolerance is always advised, especially when dealing with production workloads. You need to understand the caveats of each system and be able to plan accordingly. Containers, being more lightweight, are more resistant to failures and more readily restored.

Having understood the preceding key comparative areas, it is important to highlight the several advantages containers hold over VMs and physical machines, as follows:

• Due to better resource utilization, containers can achieve better workload density.

• They standardize deployments and simplify the application testing process.

• Application deployment is quicker and more streamlined. It is also more standardized, making it more consistent and reliable.

• Improved speed and flexibility in developing application code, as well as sharing.

You should now understand the key differences between containers and VMs, as well as be able to identify scenarios that best suit either deployment. In the next section, we will explore ACI and how this allows easy deployment of containers to the Azure platform.

ACI

ACI is a container hosting service that provides a fast and easy method for running containers in Azure. Using this solution will enable you to deploy containers without needing to manage any VMs or utilize any higher-level services. Azure container instances provide a great solution for applications that can run in isolated containers. There are several benefits to working with container instances on Azure, as outlined here:

• Fast startup times: Azure containers are designed to be able to start in seconds.

• Public connectivity: Azure containers have been designed to enable direct exposure to the internet by associating with a public Internet Protocol (IP) address that also allows the association of a fully qualified domain name (FQDN) (public Domain Name System (DNS) name associated with the public IP).

• Security: Azure container applications apply hypervisor-level security that enables container applications to be isolated as they would be in a VM.

• Custom sizes: Azure container nodes are designed to be dynamically scaled to meet the resource demands required for an application.

• Storage persistence: Azure containers enable storage persistence through Azure File Shares by allowing them to be directly mounted. Without this, storage is only ephemeral.

• Various OSs: Azure container instances can schedule containers for both Windows and Linux systems. Simply select the OS type you need when you create your container groups.

• Co-scheduled groups: Azure container instances are designed to support the sharing of host machine resources through multi-container groups on the same host.

• Virtual network (VNet) deployment: Azure container instances have been designed to allow integration into Azure VNets.

Now that you understand what Azure container instances are and the several benefits they can offer, we will start to understand how container groups fit into the structure. You will learn what container groups are and what to consider when designing and deploying these.